maintenance/storage/checkStorage.php

   1 <?php
   2 /**
   3  * Fsck for MediaWiki
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup Maintenance ExternalStorage
  22  */
  23
  24 if ( !defined( 'MEDIAWIKI' ) ) {
  25         require_once( dirname( __FILE__ ) . '/../commandLine.inc' );
  26
  27         $cs = new CheckStorage;
  28         $fix = isset( $options['fix'] );
  29         if ( isset( $args[0] ) ) {
  30                 $xml = $args[0];
  31         } else {
  32                 $xml = false;
  33         }
  34         $cs->check( $fix, $xml );
  35 }
  36
  37
  38 // ----------------------------------------------------------------------------------
  39
  40 /**
  41  * @ingroup Maintenance ExternalStorage
  42  */
  43 class CheckStorage {
  44         const CONCAT_HEADER = 'O:27:"concatenatedgziphistoryblob"';
  45         var $oldIdMap, $errors;
  46         var $dbStore = null;
  47
  48         var $errorDescriptions = array(
  49                 'restore text' => 'Damaged text, need to be restored from a backup',
  50                 'restore revision' => 'Damaged revision row, need to be restored from a backup',
  51                 'unfixable' => 'Unexpected errors with no automated fixing method',
  52                 'fixed' => 'Errors already fixed',
  53                 'fixable' => 'Errors which would already be fixed if --fix was specified',
  54         );
  55
  56         function check( $fix = false, $xml = '' ) {
  57                 $dbr = wfGetDB( DB_SLAVE );
  58                 if ( $fix ) {
  59                         print "Checking, will fix errors if possible...\n";
  60                 } else {
  61                         print "Checking...\n";
  62                 }
  63                 $maxRevId = $dbr->selectField( 'revision', 'MAX(rev_id)', false, __METHOD__ );
  64                 $chunkSize = 1000;
  65                 $flagStats = array();
  66                 $objectStats = array();
  67                 $knownFlags = array( 'external', 'gzip', 'object', 'utf-8' );
  68                 $this->errors = array(
  69                         'restore text' => array(),
  70                         'restore revision' => array(),
  71                         'unfixable' => array(),
  72                         'fixed' => array(),
  73                         'fixable' => array(),
  74                 );
  75
  76                 for ( $chunkStart = 1 ; $chunkStart < $maxRevId; $chunkStart += $chunkSize ) {
  77                         $chunkEnd = $chunkStart + $chunkSize - 1;
  78                         // print "$chunkStart of $maxRevId\n";
  79
  80                         // Fetch revision rows
  81                         $this->oldIdMap = array();
  82                         $dbr->ping();
  83                         $res = $dbr->select( 'revision', array( 'rev_id', 'rev_text_id' ),
  84                                 array( "rev_id BETWEEN $chunkStart AND $chunkEnd" ), __METHOD__ );
  85                         foreach ( $res as $row ) {
  86                                 $this->oldIdMap[$row->rev_id] = $row->rev_text_id;
  87                         }
  88                         $dbr->freeResult( $res );
  89
  90                         if ( !count( $this->oldIdMap ) ) {
  91                                 continue;
  92                         }
  93
  94                         // Fetch old_flags
  95                         $missingTextRows = array_flip( $this->oldIdMap );
  96                         $externalRevs = array();
  97                         $objectRevs = array();
  98                         $res = $dbr->select( 'text', array( 'old_id', 'old_flags' ),
  99                                 'old_id IN (' . implode( ',', $this->oldIdMap ) . ')', __METHOD__ );
 100                         foreach ( $res as $row ) {
 101                                 /**
 102                                  * @var $flags int
 103                                  */
 104                                 $flags = $row->old_flags;
 105                                 $id = $row->old_id;
 106
 107                                 // Create flagStats row if it doesn't exist
 108                                 $flagStats = $flagStats + array( $flags => 0 );
 109                                 // Increment counter
 110                                 $flagStats[$flags]++;
 111
 112                                 // Not missing
 113                                 unset( $missingTextRows[$row->old_id] );
 114
 115                                 // Check for external or object
 116                                 if ( $flags == '' ) {
 117                                         $flagArray = array();
 118                                 } else {
 119                                         $flagArray = explode( ',', $flags );
 120                                 }
 121                                 if ( in_array( 'external', $flagArray ) ) {
 122                                         $externalRevs[] = $id;
 123                                 } elseif ( in_array( 'object', $flagArray ) ) {
 124                                         $objectRevs[] = $id;
 125                                 }
 126
 127                                 // Check for unrecognised flags
 128                                 if ( $flags == '0' ) {
 129                                         // This is a known bug from 2004
 130                                         // It's safe to just erase the old_flags field
 131                                         if ( $fix ) {
 132                                                 $this->error( 'fixed', "Warning: old_flags set to 0", $id );
 133                                                 $dbw = wfGetDB( DB_MASTER );
 134                                                 $dbw->ping();
 135                                                 $dbw->update( 'text', array( 'old_flags' => '' ),
 136                                                         array( 'old_id' => $id ), __METHOD__ );
 137                                                 echo "Fixed\n";
 138                                         } else {
 139                                                 $this->error( 'fixable', "Warning: old_flags set to 0", $id );
 140                                         }
 141                                 } elseif ( count( array_diff( $flagArray, $knownFlags ) ) ) {
 142                                         $this->error( 'unfixable', "Error: invalid flags field \"$flags\"", $id );
 143                                 }
 144                         }
 145                         $dbr->freeResult( $res );
 146
 147                         // Output errors for any missing text rows
 148                         foreach ( $missingTextRows as $oldId => $revId ) {
 149                                 $this->error( 'restore revision', "Error: missing text row", $oldId );
 150                         }
 151
 152                         // Verify external revisions
 153                         $externalConcatBlobs = array();
 154                         $externalNormalBlobs = array();
 155                         if ( count( $externalRevs ) ) {
 156                                 $res = $dbr->select( 'text', array( 'old_id', 'old_flags', 'old_text' ),
 157                                         array( 'old_id IN (' . implode( ',', $externalRevs ) . ')' ), __METHOD__ );
 158                                 foreach ( $res as $row ) {
 159                                         $urlParts = explode( '://', $row->old_text, 2 );
 160                                         if ( count( $urlParts ) !== 2 || $urlParts[1] == '' ) {
 161                                                 $this->error( 'restore text', "Error: invalid URL \"{$row->old_text}\"", $row->old_id );
 162                                                 continue;
 163                                         }
 164                                         list( $proto, ) = $urlParts;
 165                                         if ( $proto != 'DB' ) {
 166                                                 $this->error( 'restore text', "Error: invalid external protocol \"$proto\"", $row->old_id );
 167                                                 continue;
 168                                         }
 169                                         $path = explode( '/', $row->old_text );
 170                                         $cluster = $path[2];
 171                                         $id = $path[3];
 172                                         if ( isset( $path[4] ) ) {
 173                                                 $externalConcatBlobs[$cluster][$id][] = $row->old_id;
 174                                         } else {
 175                                                 $externalNormalBlobs[$cluster][$id][] = $row->old_id;
 176                                         }
 177                                 }
 178                                 $dbr->freeResult( $res );
 179                         }
 180
 181                         // Check external concat blobs for the right header
 182                         $this->checkExternalConcatBlobs( $externalConcatBlobs );
 183
 184                         // Check external normal blobs for existence
 185                         if ( count( $externalNormalBlobs ) ) {
 186                                 if ( is_null( $this->dbStore ) ) {
 187                                         $this->dbStore = new ExternalStoreDB;
 188                                 }
 189                                 foreach ( $externalConcatBlobs as $cluster => $xBlobIds ) {
 190                                         $blobIds = array_keys( $xBlobIds );
 191                                         $extDb =& $this->dbStore->getSlave( $cluster );
 192                                         $blobsTable = $this->dbStore->getTable( $extDb );
 193                                         $res = $extDb->select( $blobsTable,
 194                                                 array( 'blob_id' ),
 195                                                 array( 'blob_id IN( ' . implode( ',', $blobIds ) . ')' ), __METHOD__ );
 196                                         foreach ( $res as $row ) {
 197                                                 unset( $xBlobIds[$row->blob_id] );
 198                                         }
 199                                         $extDb->freeResult( $res );
 200                                         // Print errors for missing blobs rows
 201                                         foreach ( $xBlobIds as $blobId => $oldId ) {
 202                                                 $this->error( 'restore text', "Error: missing target $blobId for one-part ES URL", $oldId );
 203                                         }
 204                                 }
 205                         }
 206
 207                         // Check local objects
 208                         $dbr->ping();
 209                         $concatBlobs = array();
 210                         $curIds = array();
 211                         if ( count( $objectRevs ) ) {
 212                                 $headerLength = 300;
 213                                 $res = $dbr->select( 'text', array( 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ),
 214                                         array( 'old_id IN (' . implode( ',', $objectRevs ) . ')' ), __METHOD__ );
 215                                 foreach ( $res as $row ) {
 216                                         $oldId = $row->old_id;
 217                                         $matches = array();
 218                                         if ( !preg_match( '/^O:(\d+):"(\w+)"/', $row->header, $matches ) ) {
 219                                                 $this->error( 'restore text', "Error: invalid object header", $oldId );
 220                                                 continue;
 221                                         }
 222
 223                                         $className = strtolower( $matches[2] );
 224                                         if ( strlen( $className ) != $matches[1] ) {
 225                                                 $this->error( 'restore text', "Error: invalid object header, wrong class name length", $oldId );
 226                                                 continue;
 227                                         }
 228
 229                                         $objectStats = $objectStats + array( $className => 0 );
 230                                         $objectStats[$className]++;
 231
 232                                         switch ( $className ) {
 233                                                 case 'concatenatedgziphistoryblob':
 234                                                         // Good
 235                                                         break;
 236                                                 case 'historyblobstub':
 237                                                 case 'historyblobcurstub':
 238                                                         if ( strlen( $row->header ) == $headerLength ) {
 239                                                                 $this->error( 'unfixable', "Error: overlong stub header", $oldId );
 240                                                                 continue;
 241                                                         }
 242                                                         $stubObj = unserialize( $row->header );
 243                                                         if ( !is_object( $stubObj ) ) {
 244                                                                 $this->error( 'restore text', "Error: unable to unserialize stub object", $oldId );
 245                                                                 continue;
 246                                                         }
 247                                                         if ( $className == 'historyblobstub' ) {
 248                                                                 $concatBlobs[$stubObj->mOldId][] = $oldId;
 249                                                         } else {
 250                                                                 $curIds[$stubObj->mCurId][] = $oldId;
 251                                                         }
 252                                                         break;
 253                                                 default:
 254                                                         $this->error( 'unfixable', "Error: unrecognised object class \"$className\"", $oldId );
 255                                         }
 256                                 }
 257                                 $dbr->freeResult( $res );
 258                         }
 259
 260                         // Check local concat blob validity
 261                         $externalConcatBlobs = array();
 262                         if ( count( $concatBlobs ) ) {
 263                                 $headerLength = 300;
 264                                 $res = $dbr->select( 'text', array( 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ),
 265                                         array( 'old_id IN (' . implode( ',', array_keys( $concatBlobs ) ) . ')' ), __METHOD__ );
 266                                 foreach ( $res as $row ) {
 267                                         $flags = explode( ',', $row->old_flags );
 268                                         if ( in_array( 'external', $flags ) ) {
 269                                                 // Concat blob is in external storage?
 270                                                 if ( in_array( 'object', $flags ) ) {
 271                                                         $urlParts = explode( '/', $row->header );
 272                                                         if ( $urlParts[0] != 'DB:' ) {
 273                                                                 $this->error( 'unfixable', "Error: unrecognised external storage type \"{$urlParts[0]}", $row->old_id );
 274                                                         } else {
 275                                                                 $cluster = $urlParts[2];
 276                                                                 $id = $urlParts[3];
 277                                                                 if ( !isset( $externalConcatBlobs[$cluster][$id] ) ) {
 278                                                                         $externalConcatBlobs[$cluster][$id] = array();
 279                                                                 }
 280                                                                 $externalConcatBlobs[$cluster][$id] = array_merge(
 281                                                                         $externalConcatBlobs[$cluster][$id], $concatBlobs[$row->old_id]
 282                                                                 );
 283                                                         }
 284                                                 } else {
 285                                                         $this->error( 'unfixable', "Error: invalid flags \"{$row->old_flags}\" on concat bulk row {$row->old_id}",
 286                                                                 $concatBlobs[$row->old_id] );
 287                                                 }
 288                                         } elseif ( strcasecmp( substr( $row->header, 0, strlen( self::CONCAT_HEADER ) ), self::CONCAT_HEADER ) ) {
 289                                                 $this->error( 'restore text', "Error: Incorrect object header for concat bulk row {$row->old_id}",
 290                                                         $concatBlobs[$row->old_id] );
 291                                         } # else good
 292
 293                                         unset( $concatBlobs[$row->old_id] );
 294                                 }
 295                                 $dbr->freeResult( $res );
 296                         }
 297
 298                         // Check targets of unresolved stubs
 299                         $this->checkExternalConcatBlobs( $externalConcatBlobs );
 300
 301                         // next chunk
 302                 }
 303
 304                 print "\n\nErrors:\n";
 305                 foreach ( $this->errors as $name => $errors ) {
 306                         if ( count( $errors ) ) {
 307                                 $description = $this->errorDescriptions[$name];
 308                                 echo "$description: " . implode( ',', array_keys( $errors ) ) . "\n";
 309                         }
 310                 }
 311
 312                 if ( count( $this->errors['restore text'] ) && $fix ) {
 313                         if ( (string)$xml !== '' ) {
 314                                 $this->restoreText( array_keys( $this->errors['restore text'] ), $xml );
 315                         } else {
 316                                 echo "Can't fix text, no XML backup specified\n";
 317                         }
 318                 }
 319
 320                 print "\nFlag statistics:\n";
 321                 $total = array_sum( $flagStats );
 322                 foreach ( $flagStats as $flag => $count ) {
 323                         printf( "%-30s %10d %5.2f%%\n", $flag, $count, $count / $total * 100 );
 324                 }
 325                 print "\nLocal object statistics:\n";
 326                 $total = array_sum( $objectStats );
 327                 foreach ( $objectStats as $className => $count ) {
 328                         printf( "%-30s %10d %5.2f%%\n", $className, $count, $count / $total * 100 );
 329                 }
 330         }
 331
 332
 333         function error( $type, $msg, $ids ) {
 334                 if ( is_array( $ids ) && count( $ids ) == 1 ) {
 335                         $ids = reset( $ids );
 336                 }
 337                 if ( is_array( $ids ) ) {
 338                         $revIds = array();
 339                         foreach ( $ids as $id ) {
 340                                 $revIds = array_merge( $revIds, array_keys( $this->oldIdMap, $id ) );
 341                         }
 342                         print "$msg in text rows " . implode( ', ', $ids ) .
 343                                 ", revisions " . implode( ', ', $revIds ) . "\n";
 344                 } else {
 345                         $id = $ids;
 346                         $revIds = array_keys( $this->oldIdMap, $id );
 347                         if ( count( $revIds ) == 1 ) {
 348                                 print "$msg in old_id $id, rev_id {$revIds[0]}\n";
 349                         } else {
 350                                 print "$msg in old_id $id, revisions " . implode( ', ', $revIds ) . "\n";
 351                         }
 352                 }
 353                 $this->errors[$type] = $this->errors[$type] + array_flip( $revIds );
 354         }
 355
 356         function checkExternalConcatBlobs( $externalConcatBlobs ) {
 357                 if ( !count( $externalConcatBlobs ) ) {
 358                         return;
 359                 }
 360
 361                 if ( is_null( $this->dbStore ) ) {
 362                         $this->dbStore = new ExternalStoreDB;
 363                 }
 364
 365                 foreach ( $externalConcatBlobs as $cluster => $oldIds ) {
 366                         $blobIds = array_keys( $oldIds );
 367                         $extDb =& $this->dbStore->getSlave( $cluster );
 368                         $blobsTable = $this->dbStore->getTable( $extDb );
 369                         $headerLength = strlen( self::CONCAT_HEADER );
 370                         $res = $extDb->select( $blobsTable,
 371                                 array( 'blob_id', "LEFT(blob_text, $headerLength) AS header" ),
 372                                 array( 'blob_id IN( ' . implode( ',', $blobIds ) . ')' ), __METHOD__ );
 373                         foreach ( $res as $row ) {
 374                                 if ( strcasecmp( $row->header, self::CONCAT_HEADER ) ) {
 375                                         $this->error( 'restore text', "Error: invalid header on target $cluster/{$row->blob_id} of two-part ES URL",
 376                                                 $oldIds[$row->blob_id] );
 377                                 }
 378                                 unset( $oldIds[$row->blob_id] );
 379
 380                         }
 381                         $extDb->freeResult( $res );
 382
 383                         // Print errors for missing blobs rows
 384                         foreach ( $oldIds as $blobId => $oldIds2 ) {
 385                                 $this->error( 'restore text', "Error: missing target $cluster/$blobId for two-part ES URL", $oldIds2 );
 386                         }
 387                 }
 388         }
 389
 390         function restoreText( $revIds, $xml ) {
 391                 global $wgDBname;
 392                 $tmpDir = wfTempDir();
 393
 394                 if ( !count( $revIds ) ) {
 395                         return;
 396                 }
 397
 398                 print "Restoring text from XML backup...\n";
 399
 400                 $revFileName = "$tmpDir/broken-revlist-$wgDBname";
 401                 $filteredXmlFileName = "$tmpDir/filtered-$wgDBname.xml";
 402
 403                 // Write revision list
 404                 if ( !file_put_contents( $revFileName, implode( "\n", $revIds ) ) ) {
 405                         echo "Error writing revision list, can't restore text\n";
 406                         return;
 407                 }
 408
 409                 // Run mwdumper
 410                 echo "Filtering XML dump...\n";
 411                 $exitStatus = 0;
 412                 passthru( 'mwdumper ' .
 413                         wfEscapeShellArg(
 414                                 "--output=file:$filteredXmlFileName",
 415                                 "--filter=revlist:$revFileName",
 416                                 $xml
 417                         ), $exitStatus
 418                 );
 419
 420                 if ( $exitStatus ) {
 421                         echo "mwdumper died with exit status $exitStatus\n";
 422                         return;
 423                 }
 424
 425                 $file = fopen( $filteredXmlFileName, 'r' );
 426                 if ( !$file ) {
 427                         echo "Unable to open filtered XML file\n";
 428                         return;
 429                 }
 430
 431                 $dbr = wfGetDB( DB_SLAVE );
 432                 $dbw = wfGetDB( DB_MASTER );
 433                 $dbr->ping();
 434                 $dbw->ping();
 435
 436                 $source = new ImportStreamSource( $file );
 437                 $importer = new WikiImporter( $source );
 438                 $importer->setRevisionCallback( array( &$this, 'importRevision' ) );
 439                 $importer->doImport();
 440         }
 441
 442         function importRevision( &$revision, &$importer ) {
 443                 $id = $revision->getID();
 444                 $text = $revision->getText();
 445                 if ( $text === '' ) {
 446                         // This is what happens if the revision was broken at the time the
 447                         // dump was made. Unfortunately, it also happens if the revision was
 448                         // legitimately blank, so there's no way to tell the difference. To
 449                         // be safe, we'll skip it and leave it broken
 450                         $id = $id ? $id : '';
 451                         echo "Revision $id is blank in the dump, may have been broken before export\n";
 452                         return;
 453                 }
 454
 455                 if ( !$id )  {
 456                         // No ID, can't import
 457                         echo "No id tag in revision, can't import\n";
 458                         return;
 459                 }
 460
 461                 // Find text row again
 462                 $dbr = wfGetDB( DB_SLAVE );
 463                 $oldId = $dbr->selectField( 'revision', 'rev_text_id', array( 'rev_id' => $id ), __METHOD__ );
 464                 if ( !$oldId ) {
 465                         echo "Missing revision row for rev_id $id\n";
 466                         return;
 467                 }
 468
 469                 // Compress the text
 470                 $flags = Revision::compressRevisionText( $text );
 471
 472                 // Update the text row
 473                 $dbw = wfGetDB( DB_MASTER );
 474                 $dbw->update( 'text',
 475                         array( 'old_flags' => $flags, 'old_text' => $text ),
 476                         array( 'old_id' => $oldId ),
 477                         __METHOD__, array( 'LIMIT' => 1 )
 478                 );
 479
 480                 // Remove it from the unfixed list and add it to the fixed list
 481                 unset( $this->errors['restore text'][$id] );
 482                 $this->errors['fixed'][$id] = true;
 483         }
 484 }